Revolutionizing Banking with Temenos

Revolutionizing Banking with Temenos

Table of Contents

1. Introduction
2. The Challenges Faced by Tomorrowbank
3. Elastic Scalability: The Solution to Unpredictable Workloads
   • Scalability Challenges
   • Bursting Out and Scaling Back
4. Ease of Maintenance and Manageability
   • The Rise of Lights-Out Data Centers
   • Leveraging Cloud Platforms for Cost Efficiency
5. Architecture Overview
   • The Role of the API Gateway
   • Routing Requests and Containerization
   • Event Processing with Kinesis and Lambda
   • Leveraging RDS and DynamoDB
6. Benefits of Using Two Databases
   • High-Speed Transaction Processing with RDS
   • Query Optimization with DynamoDB
7. Managed Services and Simplified Deployments
   • CloudFormation Scripts for Easy Deployment
   • Removing Manual Steps with Managed Services
8. Achieving Elastic Scalability
   • Scaling FarGate and Lambdas Automatically
   • Standard Metrics and Infrastructure Scaling
   • Cost Optimization with Elastic Scaling
9. Benchmarking the Architecture
   • Performance Testing and Transaction Volume
   • Ensuring Canisius Streams and Lambdas Keep Up
   • Lambda and DynamoDB Scaling with Ease
10. Conclusion

Article

Introduction

Welcome to "This Is My Architecture" with James and Tony Coleman from Tomorrowbank. In this episode, Tony discusses the challenges faced by Tomorrowbank, focusing on elastic scalability and ease of maintenance and manageability. He walks us through their architecture and explains how they leveraged managed services to address these challenges.

The Challenges Faced by Tomorrowbank

Tomorrowbank, the world's leading banking software provider, serves around 3,000 firms, including 41 of the top 50 banks. With over half a billion banking customers relying on their terminals, Tomorrowbank faces significant challenges. Tony highlights two key challenges: elastic scalability and ease of maintenance and manageability.

Elastic Scalability: The Solution to Unpredictable Workloads

The first challenge Tomorrowbank tackles is the unpredictable nature of workloads. Market-driven demands result in highly unpredictable peaks and valleys in transaction volumes. To effectively handle these workloads, Tomorrowbank required an architecture that could burst out during peak loads and Scale back down to optimize costs.

Ease of Maintenance and Manageability

The Second challenge revolves around the concept of "lights-out" data centers. Many challenger banks prefer to operate without a dedicated IT department, reducing people costs. To achieve this, Tomorrowbank leverages cloud platforms and managed services to offload maintenance and management responsibilities.

Architecture Overview

The Core of Tomorrowbank's architecture lies in the API gateway, which routes all requests, regardless of their origin. The requests are then directed to two main routes: FarGate and Query. The FarGate route utilizes an elastic load balancer to route requests to the appropriate container image, where transaction processing and data persistence occur. Event data is pushed into Kinesis and processed by Lambdas, ultimately storing optimized data models in DynamoDB.

Benefits of Using Two Databases

Tomorrowbank employs two databases, RDS and DynamoDB, to enhance performance and optimize queries. RDS serves as the low-latency transactional processing (LTP) database, designed for efficient transaction processing. To offload query workloads, Tomorrowbank adopts the CQRS pattern and leverages DynamoDB as a query-optimized database, significantly improving query efficiency.

Managed Services and Simplified Deployments

To simplify deployments, Tomorrowbank provides customers with CloudFormation scripts, enabling easy replication of their architecture. By utilizing managed services, much of the manual deployment work is eliminated, reducing the complexity and skill requirements.

Achieving Elastic Scalability

Tomorrowbank's architecture excels in achieving elastic scalability. The transaction processing engines, FarGate, and the Lambdas scale automatically to handle workload fluctuations. Leveraging the capabilities of FarGate, both container and infrastructure scaling occur seamlessly, optimizing resource allocation and costs.

Benchmarking the Architecture

In benchmarking their architecture, Tomorrowbank successfully processed over 50,000 transactions per second. By pushing the limits of Canisius streams, Lambdas, and DynamoDB, Tomorrowbank confirmed the architecture's ability to handle high transaction volumes effortlessly. The scaling capabilities of Lambdas and DynamoDB, with minimal intervention, demonstrated the efficiency and reliability of their design.

Conclusion

Tomorrowbank's architecture effectively addresses the challenges of elastic scalability and ease of maintenance and manageability. With a robust and scalable infrastructure, Tomorrowbank can handle unpredictable workloads efficiently, while leveraging managed services minimizes manual efforts and reduces costs. The benchmarking results further validate the architecture's performance and scalability, ensuring Tomorrowbank remains at the forefront of the banking software industry.

Highlights

• Tomorrowbank, the world's number one banking software provider, serves top global banks and handles transactions for over half a billion banking customers.
• The challenges Tomorrowbank focuses on are elastic scalability and ease of maintenance and manageability.
• The API gateway serves as the core of Tomorrowbank's architecture, routing all requests to either FarGate or Query routes.
• Tomorrowbank utilizes RDS for low-latency transaction processing and DynamoDB for query optimization.
• Managed services and CloudFormation scripts simplify deployments and reduce manual steps.
• Tomorrowbank's architecture achieves elastic scalability by leveraging the capabilities of FarGate and automatic scaling of Lambdas.
• Benchmarking results demonstrate the architecture's ability to handle high transaction volumes with minimal manual intervention.

FAQ

Q: How does Tomorrowbank handle unpredictable workloads? A: Tomorrowbank's architecture incorporates elastic scalability, allowing it to burst out during peak loads and scale back down to optimize costs.

Q: What databases does Tomorrowbank use and why? A: Tomorrowbank utilizes RDS for transaction processing and DynamoDB for optimized queries, resulting in high-speed processing and efficient query performance.

Q: How does Tomorrowbank simplify deployments? A: Tomorrowbank provides CloudFormation scripts that allow for easy replication of their architecture. Additionally, leveraging managed services greatly reduces the complexity and skill requirements of deployments.

Q: How does Tomorrowbank ensure scalability? A: Tomorrowbank's architecture automatically scales its transaction processing engines, FarGate, and Lambdas based on workload fluctuations, ensuring optimal resource allocation and cost efficiency.

Q: What were the benchmarking results for Tomorrowbank's architecture? A: Tomorrowbank successfully processed over 50,000 transactions per second, pushing the limits of their streaming technology and demonstrating the scalability of their Lambdas and DynamoDB with minimal manual intervention.